• KCID journal
• /
• v.17 no.2
• /
• pp.17-27
• /
• 2010

In order to manage the water quality in reservoir, it is necessary to understand the temporal and spatial variation of reservoirs and to classify the reservoirs. In this research, agricultural reservoirs are classified according to physical characteristics (depth, residence time, shape of the reservoir etc) and water quality using multivatriate analysis (PCA and CA). CA (Cluster Analysis) method classify reservoirs into several groups as a similarity of the reservoirs, but it is difficult to indicate a full list to the one table. In case of PCA (Principle Component Analysis) method, it has the advantage for the classification on the reservoirs depending on the water quality similarity and also it is useful to analyze the relationship between related factors through correlation analysis. However PCA is limited to classify into several groups based on the characteristics of the reservoirs and each user should be classified as randomly subjective according to the relative position of the reservoir in the figure. In conclusions, compared to conventional reservoirs classification methods, both CA and PCA methods are considered to be a classification method that describes the nature of the reservoir well, but classification results has a restriction on use, so further research will be needed to complement.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2003.07a
• /
• pp.259-262
• /
• 2003

The surface electronic state of rutile $TiO_2$, which is an oxide semiconductor and has a wide band gap of 3.1 $\sim$ 3.5 eV, was calculated by DV-$X_{\alpha}$ method, which is a sort of the first principle molecular orbital method and uses Hartre-Fock-Slater approximation. The $[Ti_{15}O_{56}]^{-52}$ cluster model was used for the calculation of bulk state and the $[OTi_{11}O_{34}]^{-24}$ model for the surface state calculation. After calculations, the energy level diagrams and the deformation electron density distribution map were compared in both models. As results, it was identified that the surface energy levels are found between the valence and conduction band of bulk $TiO_2$ on the surface area. The energy values of these surface-induced levels are lower than conduction band of bulk $TiO_2$ by 0.1 $\sim$ 1 eV. From this fact, it is expected that the surface energy levels act as donar levels in n-type semiconductor.

• Proceedings of the KIEE Conference
• /
• 2002.07c
• /
• pp.1515-1517
• /
• 2002

$MnO_2$ is used for the oxide electrode of electrochemical equipments because of its good electric conductivity and low oxygen overpotential. The effect of additives on the properties of $MnO_2$ has been investigated to enhance the electric conductivity and the stability in an acid solution. In this research, the effect of Ni addition on ${\beta}-MnO_2$ was studied by the theoretical quantum chemical method. The calculation was carried out by the discrete variation $X{\alpha}$ method, which is a sort of the first principle method and use Hatre-Fock-Slater approximation. The electron energy level, the density of state, the bond overlap population, the charge density distribution and the net ionic transfer between cations and anions were calculated and discussed. The used cluster model was $(Mn_{10}NiO_{44})^{-44}$.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.50 no.2
• /
• pp.142-146
• /
• 2005

Thirty-two germplasms of Korean adlay landraces were examined to analyse the genetic relationship through the amplified fragment length polymorphism (AFLP) approach. Total number of AFLP products generated by 12 selective primer combinations was 882. The number of polymorphic fragments by each primer combination greatly varied from 4 to 51 with a mean of 20.3, bands visible on the polyacrylamide gel. A genetic similarity coefficient was used for cluster analysis following UPGMA (unweighted pair grouping method of averages) method. The resulting clusters were represented in the form of a dendrogram. The clustering was not tight in the dendrogram. There was generally no clear grouping of the adlay according to the geographic regions in which germplasms were collected. The present AFLP analysis imply that although Korean adlay displayed a larger amount of AFLP variation within germplasms, the variation was shown independently without reflecting a clinal variation. This study demonstrated that AFLP method can be used to examine the genetic relationships among different germplasms of adlay.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2006.05a
• /
• pp.111-112
• /
• 2006

A new MR cylinder with built-in valves using Magneto - Rheological fluid (MR valve) is proposed for fluid power control systems. The MR fluid is a newly developed functional fluid whose obvious viscosity is controlled by the applied magnetic field intensity. This MR cylinder, which is composed of cylinder with small clearance and piston with electromagnet, has the characteristics of simple, compact and reliable structure. This paper presents a method to control the pressure of MR cylinder by using Generalized Predictive Control (GPC) algorithm. The differential pressure is controlled by applying magnetic field intensity to MR fluid. The use of GPC controller is to generate a control sequence by minimizing a cost function in such a way that the future system output is driven close to reference over finite prediction horizons. Experimental results from real time control using GPC method compared with conventional PID control method are also shown in this paper.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.7
• /
• pp.2785-2799
• /
• 2020

We propose a novel speaker adaptation method using acoustic model clustering. The similarity of different speakers is defined by the cosine distance between their i-vectors (intermediate vectors), and various efficient clustering algorithms are applied to obtain a number of speaker subsets with different characteristics. The speaker-independent model is then retrained with the training data of the individual speaker subsets grouped by the clustering results, and an unknown speech is recognized by the retrained model of the closest cluster. The proposed method is applied to a large-scale speech recognition system implemented by a hybrid hidden Markov model and deep neural network framework. An experiment was conducted to evaluate the word error rates using Resource Management database. When the proposed speaker adaptation method using i-vector based clustering was applied, the performance, as compared to that of the conventional speaker-independent speech recognition model, was improved relatively by as much as 12.2% for the conventional fully neural network, and by as much as 10.5% for the bidirectional long short-term memory.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.5
• /
• pp.2539-2554
• /
• 2017

Regression-based image super resolution (SR) methods have shown great advantage in time consumption while maintaining similar or improved quality performance compared to other learning-based methods. In this paper, we propose a novel single image SR method based on hierarchical regression to further improve the quality performance. As an improvement to other regression-based methods, we introduce a hierarchical scheme into the process of learning multiple regressors. First, training samples are grouped into different clusters according to their geometry similarity, which generates the structure layer. Then in each cluster, a compact dictionary can be learned by Sparse Coding (SC) method and the training samples can be further grouped by dictionary atoms to form the detail layer. Last, a series of projection matrixes, which anchored to dictionary atoms, can be learned by linear regression. Experiment results show that hierarchical scheme can lead to regression that is more precise. Our method achieves superior high quality results compared with several state-of-the-art methods.

• Journal of Computational Design and Engineering
• /
• v.2 no.3
• /
• pp.183-194
• /
• 2015

This paper proposes a novel method for visualizing the thickness and clearance of 3D objects in a polyhedral representation. The proposed method uses the distance field of the objects in the visualization. A parallel algorithm is developed for constructing the distance field of polyhedral objects using the GPU. The distance between a voxel and the surface polygons of the model is computed many times in the distance field construction. Similar sets of polygons are usually selected as close polygons for close voxels. By using this spatial coherence, a parallel algorithm is designed to compute the distances between a cluster of close voxels and the polygons selected by the culling operation so that the fast shared memory mechanism of the GPU can be fully utilized. The thickness/clearance of the objects is visualized by distributing points on the visible surfaces of the objects and painting them with a unique color corresponding to the thickness/clearance values at those points. A modified ray casting method is developed for computing the thickness/clearance using the distance field of the objects. A system based on these algorithms can compute the distance field of complex objects within a few minutes for most cases. After the distance field construction, thickness/clearance visualization at a near interactive rate is achieved.

• Korean Journal of Crystallography
• /
• v.14 no.1
• /
• pp.16-23
• /
• 2003

The electronic structure and chemical bonding of β-MnO₂ were theoretically investigated by DV-X/sub α/ (the discrete variation X/sub α/) method. which is a sort of the first principle molecular orbital method using Hatre-Fock-Slater approximation. The calculations on several cluster models having different sizes were carried out for the determination of a model suited for analyzing bulk state. The Mn/sub 15/O/sub 56/ model was selected as a sufficiently suitable model for the calculation of electronic state and chemical bonding by the comparison of the calculated XPS (X-ray photo-electron spectrum) and experimentally measured XPS. By using this model, the electron energy level, the density of state, the bond overlap population, the charge density distribution, and the net ionic transfer between cations and anions were calculated and discussed.

• Journal of the Korean housing association
• /
• v.25 no.4
• /
• pp.59-72
• /
• 2014

This study proposes a data-mining analysis method for examining variable multi-attribute big-data, which is considered to be more applicable in social science using a Correspondence Analysis of variables obtained by AIC model selection. The proposed method was applied on the Seoul Survey from 2005 to 2010 in order to extract interesting rules or patterns on characteristics of single household. The results found as follows. Firstly, this paper illustrated that the proposed method is efficiently able to apply on a big dataset of huge categorical multi attributes variables. Secondly, as a result of Seoul Survey analysis, it has been found that the more dissatisfied with residential environment the higher tendency of residential mobility in single household. Thirdly, it turned out that there are three types of single households based on the characteristics of their demographic characteristics, and it was different from recognition of home and partner of counselling by the three types of single households. Fourthly, this paper extracted eight significant variables with a spatial aggregated dataset which are highly correlated to the ratio of occupancy of single household in 25 Seoul Municipals, and to conclude, it investigated the relation between spatial distribution of single households and their demographic statistics based on the six divided groups obtained by Cluster Analysis.